Nearest neighbor classifiers recognize stroke gestures by computing a (dis)similarity between a candidate gesture and a training set based on points, which may require normalization, resampling, and rotation to a reference before processing. To ...